A systematic review of the effectiveness, safety, and acceptability of medical management of intrauterine fetal death at 14–28 weeks of gestation

Optimal dose, interval, and administration route of misoprostol with added benefit of mifepristone for management of second trimester intrauterine fetal death (IUFD) are not established.

Active rather than expectant management is often favored because of the increased risk to the mother, secondary to retained products of conception. 1 Additional concerns for the woman include the risk of uterine rupture and disseminated intravascular coagulation. 1 Medical management is considered a safe and effective treatment option for IUFD. 4 One common medication used to treat IUFD is misoprostol, 1 which is a synthetic prostaglandin E1 analogue.
Misoprostol is effective in emptying the uterus owing to its ability to induce uterine contractions and soften the cervix. 4 Adverse effects are often mild and include nausea, vomiting, diarrhea, fever, chills, bleeding, and pain due to uterine contractions. 4 Several misoprostol regimens exist, including use alone or with other agents such as mifepristone. Studies evaluating the safety, effectiveness, and acceptability of the drug alone and/or in combination with mifepristone have often included viable pregnancies with induced abortion or labor induction. 5 Previous systematic reviews including IUFD have not focused exclusively on the second trimester or consistently defined IUFD; as such, limited guidance has been provided for pregnancies at 14-28 weeks of gestation. [6][7][8] Questions pertaining to the optimal dose, interval, and route of administration of misoprostol, as well as the added benefit of mifepristone for management of second trimester IUFD still remain. The aim of the present systematic review is to address these knowledge gaps by assessing the effectiveness, safety, and acceptability of medical management of second trimester IUFD.
This review was conducted as part of the evidence base for World Health Organization (WHO) recommendations on medical management of abortion.

| MATERIALS AND METHODS
At the outset of this review, the most recent systematic review focusing on medical management of IUFD included randomized controlled trials and was conducted in 2006. 8 We sought to update this review focusing on medical management using mifepristone and/or misoprostol for pregnancies at 14-28 weeks of gestation. Appendix S1) was developed together with experts working in systematic review literature searching at WHO, Geneva, and Karolinska Institutet, Stockholm. Our search strategy was formed by combining two concepts: (1) IUFD; and (2) misoprostol. The key words used were "fetal death"; "abortion, missed"; and "misoprostol" to capture studies using various terms to describe IUFD, and those that used misoprostol alone and a combination of mifepristone and misoprostol (combined regimen). MeSH terms were used when applicable. The search strategy was customized to each electronic database with no language restrictions.
We hand searched the reference lists from three systematic reviews that included second trimester IUFD, [6][7][8] and the reference lists of eligible trials identified through our search. There was no patient or public involvement in the development of this review.
We included randomized controlled trials with cases of IUFD at between 14 and 28 weeks of gestation, where cases were evenly distributed between study arms. Trials including IUFD below 14 weeks or above 28 weeks of gestation were only considered if the mean gestational age of the participants was within 14-28 weeks. Trials were selected for inclusion if they included comparisons of misoprostol alone or a combined regimen using different routes or dosages of misoprostol. Trials comparing misoprostol alone or a combined regimen with expectant care, placebo, or surgery were also included.
Two reviewers conducted the screening, data extraction, and assessment of risk of bias, in parallel. Standardized screening and data extraction forms were created prior to data collection.
Titles and abstracts were screened, and full text was obtained if both reviewers judged a citation to be potentially eligible. Risk of bias was assessed in included studies in accordance with the Cochrane Handbook. 9 Selection, performance, detection, attrition, and reporting bias were considered key domains in the summary assessment of risk of bias within each trial (Table S1). 10 Any discrepancies were reviewed and discussed between the authors and resolved together.
We included trials where the primary effectiveness outcome was complete abortion, defined in this review as complete expulsion of products of conception, measured within 24 or 48 hours (utilizing a variety of assessment methods, including ultrasound or clinical signs and symptoms). Other effectiveness outcomes were induction to expulsion interval (time from treatment initiation to complete abortion) and need for surgical intervention. Our safety outcome was measured as number of serious adverse events (SAEs; such as hospitalization, blood transfusion, need for further surgery beyond interventions to complete removal of products, or death). Secondary outcomes included adverse effects (including nausea, vomiting, fever, pain, and hemorrhage) and acceptability (measured in satisfaction). A core outcome set related to abortion exists (http://www. crown-initiative.org/core-outcome-sets/ongoing-core-outcomesets/); however, since it is still in development it has not been used in this review.
We used RevMan as our analysis tool. 11 For dichotomous outcomes we analyzed data based on number of events and number of women assessed in the intervention and comparison groups. We used these to calculate the relative risk (RR) and 95% confidence interval (CI). For continuous outcomes we reported the measure as mean ± SD. We used an online converter (http://vassarstats.net/ median_range.html) and the recalculation method proposed by the Cochrane Training group (http://train ing.cochr ane.org/resou rce/analy sing-conti nuous-outcomes). The certainty of the evidence using the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) 12 were assessed and presented alongside the findings in the tables.
As the present systematic review is based on existing published literature, no ethical clearance was required.

| RESULTS
The initial search generated 1695 citations after duplicates were removed. Eighty-three full texts were screened, of which 67 were excluded. Six additional trials [13][14][15][16][17][18] identified from three systematic reviews 6-8 were included. We contacted the authors of included trials where additional information was needed. We received information from four trial authors, [19][20][21][22] with two providing disaggregated IUFD data. 19,20 The flow of included and excluded trials is presented in Figure 1.
The characteristics of the included trials are presented in Table 1. We included 16 trials with a total of 1890 participants. Four trials 14,15,19,23 included women with a single prior uterine incision; one trial 21 stated that women with a prior uterine incision were eligible but it is unclear how many were included. No direct evidence related to medical management compared with surgical, expectant, or placebo management was identified. Four trials reported that the analysis was intention-to-treat (ITT) analysis. 18,20,23,24 Thus, it was not possible to perform true ITT analyses for all comparisons and outcomes.
One trial compared mifepristone and misoprostol versus misoprostol alone. 20 Misoprostol was administered vaginally every 6 hours.
Women treated with the combined regimen had slightly higher rates of complete abortion within 24 hours (RR 1.18; 95% CI, 0.91-1.53) and  a shorter expulsion time than women treated with misoprostol alone.
No SAEs were reported among 27 participants. The evidence is very low for all reported outcomes (Table 2A). The effect estimates for all the outcomes were based on ITT analysis.
Two trials compared the sublingual route with the oral route. 15,25 In these trials a dose of 100 μg misoprostol was administered every 2 and 4 hours, respectively. The sublingual route was more effective when considering completeness (RR 0.88; 95% CI, 0.70-1.11) and The confidence interval (95% CI) around the effect estimate indicates both a positive and negative effect. b GRADE explanation: Downgraded two levels in certainty due to serious imprecision. c GRADE explanation: Downgraded three levels in certainty due to very serious imprecision.
T A B L E 5 Comparison of route of misoprostol when misoprostol is used alone.  This systematic review also has limitations. The certainty of the evidence was in general low or very low for many outcomes, mainly owing to the existence of few trials with relatively small sample sizes and unclear risk of bias. The trials showed statistical heterogeneity as well as imprecision, and outcomes were defined differently across trials.

| DISCUSSION
Additionally, few trials reported on IUFD separately or provided disaggregated data upon request making the actual sample size of IUFD cases small owing to mixed study populations. Nevertheless, few systematic reviews have focused on medical management of second trimester IUFD and this review may provide valuable information with regard to management and future research needs.
WHO recommends a combination of mifepristone and misoprostol for the management of first and second trimester induced abortion in viable pregnancies. 29 Findings from a recently updated review on the management of fetal death below 24 weeks by Lemmers et al., 30 suggest that the mifepristone-misoprostol regimen is equally effective to misoprostol alone. However, the data on which these findings are based do not include any cases of second trimester IUFD. In the present review, the certainty of evidence concerning combined regimens was very low, hence we are uncertain about its benefits in the management of second trimester IUFD. Our analyses suggest that there may be a small advantage of adding mifepristone 36-48 hours prior to misoprostol, but further research is needed to establish this effect. Nevertheless, the effectiveness demonstrated for combined regimens when used for other indications, 5 together with our analysis, suggests that the advantage of a combined regimen also may apply to second trimester IUFD.
When mifepristone is not available or feasible, misoprostol can be used alone. 29 When misoprostol is used alone our analyses suggest that 400 μg misoprostol is more effective compared with lower doses and that women receiving 400 μg experience fewer adverse effects. When comparing 400 μg with 600 μg, we found that there may be little to no difference in completeness, but that the higher dose results in a slightly shorter expulsion interval. Our analysis suggests that the 600μg dose also leads to more adverse effects; however, we are uncertain about the effect on adverse effects for this comparison owing to very low certainty evidence.
Thus, the 400μg dose may be the lowest dosage at which effectiveness is relatively high and the rate of adverse effects relatively low.
Our results indicate that there may be an advantage to using a loading dose of misoprostol followed by 200 μg misoprostol compared with 200 μg misoprostol alone. However, when comparing a loading dose followed by 200 μg misoprostol with 400 μg misoprostol alone, differences in effectiveness were diminished. While we may not know the true effect regarding adverse effects owing to limited certainty of evidence, our analyses suggest adverse effects also improve when the loading dose is not administered, rendering the loading dose unnecessary.
Evidence extrapolated from studies on second trimester induced abortion suggests that a shorter interval of misoprostol dosing of 3 hours compared with 6 hours is more effective and does not compromise safety. 5 Included trials in our review used treatment intervals ranging from 2 to 12 hours but no trials compared timing of misoprostol. The dose of 400 μg, which we found to be superior in terms of effectiveness and adverse effects, was administered every 4 hours.
Although this may be a reasonable interval, research is needed to establish whether this interval is ideal.
While effectiveness data may vary across routes, women value choice as it relates to their abortion experience. 29  The uterus becomes more sensitive to prostaglandins with increasing gestational age, 4 but reports of severe complications such as uterine rupture after use of misoprostol are extremely rare. 31 This review included 60 women with prior uterine incision out of a total of 1890 women and no cases of uterine rupture were reported. Eight SAEs were reported by two trials; however, most trials reported no SAEs or did not report on safety at all. Included trials were also not powered to detect group differences in safety. Although SAEs were rare, and medical abortion medications are generally safe, 4,29,31 we cannot conclude on safety in this review due to limited certainty of evidence. Furthermore, our data were limited in gestations above 24 weeks; thus, generalization of our findings to IUFD above 24 weeks should be made with caution.
Our primary outcomes were effectiveness measured by completeness, expulsion time, and need for surgical intervention. It is important to understand what value women place on such outcomes. While most of the studies included did not report on acceptability, Bracken et al. 24 provide key insights into women's preferences; women preferred a higher dose of misoprostol, despite increasing adverse effects, possibly due to greater effectiveness. More information is needed to better understand the trade-offs considered and made by women, especially as they relate to these outcomes and the existence of adverse effects.
In conclusion, this review provides information related to the dose, route, and preparation of misoprostol in misoprostol-only regimens. We are uncertain about the added benefit of mifepristone to misoprostol. Our findings suggest that a regimen of 400 μg misoprostol every 4 hours, administered sublingually or vaginally, may be effective in the management of second trimester IUFD. We are unable to draw any conclusions regarding safety and acceptability owing to limited evidence.

AUTHOR CONTRIBUTIONS
AC and AL conceived the idea and conducted the search, screening, data extraction, and quality assessments. MSF conducted the analyses and GRADE. All authors carried out the interpretation of the data, revised the article, and approved the final version for publication.

CONFLICTS OF INTEREST
The authors have no conflicts of interest.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Reported side effects and satisfaction -Comparison of different misoprostol-mifepristone regimens.   Appendix S1. Search strategy (as searched in Pubmed).